Posts Tagged ‘Software’

Is SaaS cheaper than licensed software?

November 21, 2008

This is a blog post from ReadWriteWeb, where I occasionally contribute.

Most people quickly answer this question in the affirmative; I certainly do.  However, there are people out there who aren’t sure.  They look at the monthly cost of a SaaS application and compare it to an equivalent licensed product over an extended period of time.  Given enough time, you will eventually hit a point where the SaaS product “appears” more expensive.  Let’s look at it from total cost of ownership (TCO) perspective.

The true cost of a licensed product is MUCH higher than just the software.  Here are other things to factor in:

  • Hardware costs – You either have to buy machines or add your software to existing servers and manage them.  If it is a mission critical application, you will probably need dedicated machines and back-ups
  • Additional software costs – You will most likely need an OS, application server software, database, monitoring software, etc.. Many of these products are OpenSource now, but there are still associated costs
  • Implementation costs – In my experience, the implementation costs associated with a behind the firewall solution are ALWAYS higher than a SaaS application.  There is simply more to do.  You will either pay consultants or use your own valuable resource time to worry about installing software, integrating it, building servers, configuration, etc..
  • Maintenance labor – If you have software in-house, there is going to be some level of effort required to keep it happy.  Your IT resources will need to take care of it, which will keep them from doing more value-added activities.

Another huge factor here is the ability to get the latest and greatest technology.  Once you install software in a data center, it becomes more difficult to upgrade and maintain it (especially if you customize it).  In this case, you will be stuck with old software that you will have to replace in the same time horizon as we described above.  In other words, unless you are absolutely sure beyond a shadow of the doubt that your licensed software is going to meet your business needs for 5 years or longer, it might make financial sense.

Let’s look at a real-world example.  There is a 100 person company that has been sharing files via email and internal servers.  The executives have finally concluded they need to join the 21st century and a solution in place.  One option is to implement SharePoint.  Here is a rough estimate of what that might cost:

Year 1
MOSS Server – $4,500
User Client Access License – $90
Hosting and Maintenance – $5,000
Implementation/Developer Support – $20,000
Total – $29,590

Year 2 – X
Hosting and Maintenance – $5,000
Developer Support – $3,000
Total – $8,000

I know of a SaaS solution that has 80% of the file collaboration functionality of SharePoint that charges $850 per month for 100 users.

Year 1
SaaS Fees – $10,200
Implementation Support – $10,000
Total – $20,200

Year 2 – X
SaaS Fees – $10,200
Total – $10,200

It will take over 4 and a half years before the license software is cheaper.  By that time, I’m quite sure there will be another solution that replaces SharePoint and the cycle starts again.  We can dither about the numbers but you get the point.  Plus, the numbers don’t reflect that the SaaS solution is likely to improve and innovate faster than the licensed software by a significant amount.

What do you think?  Have you done this analysis and what did you conclude?

Quality Assurance (i.e. testing) Is A Roller Coaster

November 14, 2008

We are in the heat of preparing a new release, which will be let out into the wild next week.  The ups and downs of building a release are quite dramatic.  I’ll focus on testing for this conversation.  It is a funny set of emotions.  In the beginning, you feel like kid opening presents for your birthday.  After talking about new features for weeks, designing them, arguing about how the work, etc, you get your first glimpse of the actual new application and features.  This part is really fun.

The next phase is when you get into testing.  In the beginning, you find bugs but you get in a groove.  The bugs are usually in the new stuff, so they are relatively easy to find.  This part is good because you feel like you are really working.

The phase after this one is the toughest (the one we are in right now).  The application is really shaping up, and it is much tougher to find bugs.  On one hand, this is obviously a really good thing.  For the tester, it is really boring and tedious.  You search for things that are wrong and try to think of as many edge cases as you can.  Every now and then, you get rewarded.  As the days go by, it becomes harder to break things (again, a good thing).

Finally, you finish the testing phase and get ready to release the software.  The excitement builds.  We put up system notices letting our customers know.  Then we go to sleep knowing there is a shiny new present that will be waiting the next day.  In the middle of the night, the boys in Sweden do their magic and presto.  This is the REALLY fun and exciting part.  Hope this helps explain one small part of the software development process.

Focus on user adoption, not software features

November 12, 2008

This is a blog post from ReadWriteWeb, where I occasionally contribute.

I sat through a very interesting presentation at the OpenAir User Conference. The key takeaway was a statistic on achieving enterprise software success.  Contrary to most of what we cover on blogs, marketing, demos, etc., effective user adoption is the absolute best predictor of enterprise software success.

According to a study done by the Sand Hill Group and Neochange, the most critical factor (70% listed as number 1) for software success and return-on-investment is effective user adoption. Software functionality came in at 1% surprisingly, with organization change at 16% and process alignment at 13%. This is a remarkable result. You can have the best software in the world, with the most sophisticated features, analytics and integration, blah blah blah, but if people don’t use it, it isn’t going to add value. I can’t tell you how many RFPs and software selection processes I’ve been involved with in prior lives that focus almost exclusively on tiny little, “knat’s ass” features that few people if at all will ever use. This study shows that focusing so much on features is missing the boat entirely.

This finding is very interesting for all kinds of applications, including consumer applications.  Features very rarely make someone take to an application or not.  Moreover, I doubt most software companies really take user adoption as a holistic approach into account when designing their applications.  If this trend is accurate (and my experience tells me it is), then I think it has very interesting ramifications on how software should be designed, sold and implemented.  User adoption is typically something that comes at the end of a cycle.  This says it should be one of the most important elements of the entire process.  Please share any opinions or war stories that either confirm or refute this conclusion.

User Adoption Key for Software ROI

October 15, 2008

I sat through a very interesting presentation yesterday.  The topic was how the software industry has fundamentally changed and that services is a growing part of software company revenue, whether they like it or not.  Customers expect solutions these days and aren’t interested in having a disk shipped to them and wishes for good luck.  While this is interesting, the key takeaway for me was a statistic on achieving enterprise software success.

According to a study done by the Sand Hill Group and Neochange, the most critical factor (70% listed as number 1) for software success and return-on-investment is effective user adoption.  Software functionality came in at 1% surprisingly, with organization change at 16% and process alignment at 13%.  This is a remarkable result.  You can have the best software in the world, with the most sophisticated features, analytics and integration, blah blah blah, but if people don’t use it, it isn’t going to add value.  I can’t tell you how many RFPs and software selection processes I’ve been involved with in prior lives that focus almost exclusively on tiny little, “knat’s ass” features that few people if at all will ever use.  This study shows that focusing so much on features is missing the boat entirely.

There are several factors contributing to effective user adoption that include:

  • Training
  • Change management strategies and tactics
  • Executive support
  • Software ease-of-use
  • Etc.

I’m very encouraged by this as a predictor for GroupSwim’s success.  We have built an incredibly easy-to-use application that is fun to use.  Our customers and analysts have told us this time and time again.  It is a fact that GroupSwim is easy to use and requires no training.  If you believe this, we have half the requirements covered without spending a dime.  If our customers and prospects use the application to change how they work or interact with their customers and partners, they get LOTS of value while expending very little effort or capital.  I hope to see more studies like this in the future :)

Executing a Major Software Release

September 11, 2008

As you probably know, we are gearing up for a major release of GroupSwim next week.  I thought you might be interested in the things we do to prepare for this event.

  1. Testing, Testing, Testing – Did I mention we test?  In all seriousness, it is the most important component of the process.  Quality and customer success are critical, so we try to avoid any and all bugs.  We also learn about the software by everyone pitching in on the testing process.  We often tweak designs and the user interface based on what we learn.  As good as our initial designs are, nothing trumps rolling up the sleeves and playing with the software.  As someone who sells and demos GroupSwim every day, I find there is no better way to learn it.  The engineering team takes the bugs we find and turns-around fixes overnight.  We are a well-oiled machine, but it takes time and effort.
  2. Technical Tasks – Rolling out a major release involves important technical tasks.  Once we take the system off-line to do the migration, we perform back-ups, checks, software migrations, queue-up requests, redirect URLs, and a host of other activities.  Since GroupSwim is a SaaS application, it requires a different set of tasks (and mindset) than traditional software.  We aren’t migrating one customer; we are migrating 1,000s. The engineering team does an amazing job in a very short time frame.  We do mock go-lives and test with copies of production data to make sure we are as rigorous as possible.
  3. Training and Help – When we roll-out software of this magnitude, we update everything.  The list of stuff we change includes but is not limited to:
    • The website.  This particular release will touch almost every page
    • Marketing collateral
    • Online help.  This is a big task when we add so many features
    • Seed content for new sites.  These are the discussions, documents and wikis we put in every new site a customer creates to give them a running start
    • Videos
    • Update the Pool.  This is our own customer community, so we put announcements, tips and tricks and other content on it
    • Demo sites
  4. Marketing Activities – When we create software of this magnitude, we want to tell the world.  Marketing takes a significant amount of effort.  For starters, we do the following:
    • Write and distribute press releases
    • Email past and present users
    • Reach out to bloggers and analysts
    • Email campaigns to leads and contacts

Every time we release software, we learn something new.  I’m giddy about this next one and look forward to a new stage in our company’s evolution.

Christmas in September – We opened our present this morning

September 2, 2008

For those of you who haven’t worked at a software company (nice move by the way), the first look at a new software release brings a palpable feeling of excitement. Our engineering team just released the first Test Candidate for our next release.  It contains the new wiki and a host of other new features. It is very exciting to see our wiki in reality after talking about it and planning it for the past several months. I can’t wait to see it out in the market and our customers using it.

The testing process is both exciting and boring at the same time.  For the next 2 weeks, we will hammer the system as hard as we can.  We’ll try to do everything we can to anticipate how users will use the system.  We’ll test the “happy path” (when people use the software they way we anticipated) and more unconventional use cases that users will undoubtedly try. Each day we go through as many iterations and use cases as we can, and file any bugs we find.

The engineers in Stockholm and here then fix the bugs we find as we sleep, make improvements, and release a new test candidate for us to start all over on the following day.  Starting next week, we’ll issue Release Candidates which will be closer to General Availability in quality.  We’ll do several of those and ultimately release the software in about 2 weeks.

This process is very similar to what other software companies use, but we have it down to a science. Wish us Happy Testing!

Implementing Collaborative Software – No Problem

July 2, 2008

I’ve been thinking about the difference between implementing a collaboration application for business versus more traditional applications like ERP, CRM, etc. (insert other acronym here).  I’ll try not to be biased but I believe that implementing an application like GroupSwim is different and easier.  First, let’s list a couple of assumptions and my conclusions:

Assumption #1: The amount of archived/older information that people rely on for their day-to-day knowledge-based jobs is relatively small

Conclusion #1: The need for collaboration software to integrate with older documents and information stores is minor but often overblown.  I know this flies in the face of tradition, but I think it is true.  From what I’ve experienced, people use a small subset of documents, resources, etc. and do most of their work from their inbox and on specific documents that are part of their current plate of work.  This isn’t necessarily a good thing but is reality; people aren’t accessing 1 year old plus documents very often.  For example, if you are product manager, you probably do most of your work on current product specs, analysis, customer focus groups, etc..  If you are a sales person, you use a small set of documents (marketing collateral, emails you like to send, account plans, contracts, etc.) to find and close deals.  The list goes on but you get it.  So, I don’t think making every single document every created available to a new system is important.  There is no doubt that there is a group of documents (training, in process files, etc.) that are important, but they can be added manually or integrated without trying to integrate everything.

Assumption #2: Collaborative work is dynamic and constantly changing

Conclusion #2: There are few examples of a “best” time to roll-out a collaborative solution.  Employees are constantly collaborating on something so just get started.  There are reasonable examples of where it makes sense.  For example, if a group get reorganized, or a new project or process gets kicked off, etc., these are good times to start using a collaborative solution like GroupSwim.  However, my conclusion is that now is as good a time as any to get started.  I think that most “knowledge” work requires collaboration on an every day basis, so there is no time like the present to bite the bullet and get started.

Assumption #3: Collaboration is an ill-defined process at best

Conclusion #3: The change management and leadership required to make people collaborate differently is even more onerous than traditional systems.  For example, when you implement a new CRM, ERP or whatever, your work process to get things done “must” change.  You are essentially forced to modify the way you work whether you like or not based on the new system you must use.  With a system like GroupSwim or other collaborative solutions, you can still use email and other means to get work done even though the outcome and process might suck.  So, it is incumbent on a champion and other internal leaders to make sure people buy-in and use new collaborative tools like GroupSwim to change the they work.  Eventually, they will get it and not require the oversight but it is critically important in the beginning to change habits.

Taking all three points into consideration, I think the collaboration implementations generally overestimate the need to integrate and deal with technical issues and underestimate the need for leadership and change management.  It is always harder to change people’s work habits and soft skills versus changing a discrete process or operational system.  What do you think?